Our Methodology
Last updated: August 2026
This page explains how CleanBotPick arrives at a recommendation, where the information comes from, and what the CleanScore™ does and does not tell you. If a score is going to sit at the top of an article, you are entitled to know how it was produced.
Who Does This Work
CleanBotPick is a one-person site. There is no test lab and no editorial staff. Articles are published under the site name rather than a personal byline because the site, not an individual expert, is what stands behind the analysis. More about the operation is on the about page.
What We Look At
- Manufacturer specification sheets, manuals, and support documentation.
- Retailer listings, for the current configuration and price range of a product.
- Published third-party test results and certification data, where a manufacturer or an independent organization has made them public.
- Documented owner feedback -- public forum threads, community posts, and manufacturer Q&A -- for the failure modes that specification sheets never mention.
What We Don't Do
- We do not test products. We have no lab, no instruments, and no sample units. Nothing on this site is a first-hand performance measurement.
- We do not accept free review units or sponsorships. No manufacturer can buy a position in a ranking, and none has ever been offered one.
- We do not describe experiences we have not had. If an article says a product runs quietly, that traces back to a published figure or to an owner who said so in public, and the article says which.
The site is funded by affiliate commissions on purchases made through our links, at no extra cost to you. That funding model is described in full in our affiliate disclosure. It pays for the research; it does not decide the outcome.
How the CleanScore™ Is Calculated
The score is a weighted average of published specifications, expressed on a 0–100 scale. It is arithmetic, not judgment, and it runs the same way for every product in a category:
- Record the published specifications for each product from the sources listed above.
- Normalize each specification onto a 0–100 sub-score against a fixed reference range for that category, so the same figure always produces the same sub-score.
- Combine the sub-scores as a weighted average.
- Where a specification is not published, drop that dimension and re-weight the remainder. A product is not penalized for a number its maker chose not to disclose -- but it is then scored on less evidence.
- Round to a whole number, and store it so the same product carries the same score in every article on this site.
Every scorecard on this site names the factors behind its numbers. For this category, the CleanScore™ weighs the following, and each factor is drawn from published data rather than from anything we did ourselves:
| Dimension | What it is drawn from |
|---|---|
| Cleaning performance | rated suction in pascals, brush design, and the surfaces the maker says it handles |
| Navigation and mapping | the sensing hardware the manufacturer documents, and what mapping features come with it |
| Mopping capability | tank size, pad type, and whether mopping is a real mode or a bolt-on |
| Battery and auto-empty | rated runtime, recharge-and-resume behavior, and what the dock does on its own |
| Value per dollar | everything above weighed against the product's typical price range, including brushes, filters, and bags |
Scores map to grades on fixed bands:
| Score | Grade |
|---|---|
| 85–100 | Excellent |
| 70–84 | Great |
| 55–69 | Good |
| 40–54 | Fair |
| Below 40 | Below Average |
Where the Score Falls Short
A number carries more authority than it has earned unless its limits are stated, so here are ours.
- The score reflects what a manufacturer published, not what a product did. It is not a measurement.
- A product scored on three dimensions is a weaker result than one scored on five, and the score itself does not show you which you are looking at.
- Manufacturers do not all measure the same way, so headline specifications compare poorly across brands.
- Suction figures in pascals are self-reported and are not measured the same way by every manufacturer, so they compare poorly across brands.
- Prices move constantly, and any score with a value component drifts as they do.
Treat the score as a way to build a shortlist quickly, not as a verdict. For anything expensive, read the comparison table and the caveats underneath it.
Updates and Corrections
Specifications change, products are discontinued, and we get things wrong. When an article is revised, the updated date at the top of it changes. If you find an error -- a wrong specification, a stale price range, a product that no longer exists -- tell us and we will correct it and say that we did.